Results from a mailed promotion of medication reviews among Department of Defense beneficiaries receiving 10 or more chronic medications.

BACKGROUND: Like many payers, the Department of Defense (DoD) has limited ability to work directly with prescribers to ensure appropriate medication use. Many older beneficiaries are prescribed multiple maintenance medications, placing them at higher risk for adverse drug interactions, contraindicated medication use, and other polypharmacy-related problems. Medication reviews may mitigate these risks, but the optimal venue for medication therapy management is unclear.

OBJECTIVES: To (a) determine if beneficiaries will respond to a mailed request from the DoD to pursue a medication review; (b) identify medication review location and outcomes from the patient perspective; and (c) assess the statistical significance of changes in the number of prescription medications overall and for key categories, including maintenance medications and contraindicated medications, relative to a propensity-matched comparison group.

METHODS: A total of 4,000 TRICARE beneficiaries aged 55 years or older, residing in North Carolina, who obtained 10 or more maintenance medications (defined by a unique combination of drug, strength, and dosage form) during the 90-day baseline period from May 3, 2008, to July 31, 2008, were mailed letters requesting their participation in the study. Consenting subjects received a personalized medication list to review with their physicians or pharmacists and a survey form to complete after the review. Survey results were compared by location of medication review (i.e., physician's office, pharmacy, or both). Changes from the 90-day baseline to 90-day post-intervention period were calculated for prescription utilization measures (total drug count, maintenance drug count, count of Beers list medications, and count of contraindicated drug combinations) for the subsample of subjects who completed the survey (n = 373) and for subjects who received the initial consent letter (n = 3,856) versus a propensity-matched comparison sample drawn from neighboring states. Variables included in the propensity score were gender, age group, military rank, catchment status indicating proximity to military pharmacies, enrollment status, number of pharmacy settings used, and each of 30 binary disease indicators.

RESULTS: A total of 1,469 subjects responded to the consent letter (response rate = 38.1%); 606 subjects consented to participate (consent rate = 15.7%); and 373 subjects returned a completed survey (completion rate = 9.7%). Among those who completed the survey, 190 (50.9%) received reviews in a physician's office; 103 (27.6%) received reviews in a pharmacy; 60 (16.1%) received reviews in both locations; and 20 (5.4%) reported a different location or no location. 61 survey respondents (16.4%) indicated that they were told to stop a medication, and 77 (20.6%) reported a dosage change. Medication changes occurred significantly more frequently for reviews performed at a physician's office compared with other review locations. Therapeutic classes most frequently stopped or adjusted for dosage were antidiabetics, diuretics, antilipidemics, renin-angiotensin aldosterone system inhibitors, anticoagulants, nonsteroidal anti-inflammatory drugs, and beta-adrenergic blocking agents. 85% of respondents reported that the medication review was worth doing. In the assessments of changes in prescription utilization from the baseline to post-intervention periods, no significant by-group differences were noted among those who completed the study relative to their matched comparison subjects. In the comparison of subjects who received the initial consent letter with their matched counterparts, small but statistically significant differences were observed for several prescription utilization measures, including changes in use of high-risk Beers list medications (P = 0.033); use of electrolytic, caloric, and water balance medications (P = 0.038); and use of hypertension medications (P = 0.028). The magnitude of the decrease observed among comparison subjects, however, exceeded that observed among the case subjects.

CONCLUSIONS: Response was poor to a mailing that promoted a beneficiary- initiated medication review. The absence of significant changes following the medication review suggests several possibilities: a mailed intervention is ineffective in promoting medication review; medication regimens for study subjects are already optimized to the extent obtainable through a routine medication review; or the study sample size was too small to detect relevant changes. Most drug regimen changes were dosage adjustments for current medications or substitutions within the same therapeutic class. The extent to which comprehensive assessment of a patient's medication regimen, including nonprescription and herbal agents, was performed is unclear. More intensive interventions may be required to ensure that medication regimens are being actively managed among those who use a large number of prescription medications.